Zobrazeno 1 - 10
of 35
pro vyhledávání: '"Pawel Rozycki"'
Autor:
Roman Peleshchak, Pawel Rozycki, Mariusz Wrzesien, Ivan Peleshchak, Vasyl Lytvyn, Janusz Kolbusz, Jan Kopka, Janusz Korniak
Publikováno v:
Advances in Intelligent Systems and Computing ISBN: 9783030511852
This paper shows a new type of artificial neural network with dynamic (oscillatory) neurons that have natural frequencies. Artificial neural network in the mode of information resonance implements a new method of recognition of multispectral images.
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::6a5f4f862924a237cd326746724fa0bf
https://doi.org/10.1007/978-3-030-51186-9_22
https://doi.org/10.1007/978-3-030-51186-9_22
Publikováno v:
IEEE Transactions on Industrial Informatics. 14:931-940
Radial basis function (RBF) networks, because of their universal approximation ability, have been widely applied to industrial process modeling. In this study, an Improved ErrCor (IErrCor) algorithm—an extension of error correction (ErrCor) algorit
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::6968251fc5b3780af7d8160bd9c9a59f
https://doi.org/10.1109/tii.2017.2734686
https://doi.org/10.1109/tii.2017.2734686
Publikováno v:
HSI
National Information Processing Institute
National Information Processing Institute
RBF networks seem to be an interesting and efficient alternative for traditional sigmoid-based neural networks. More sophisticated activation function makes a network more powerful but requires developing of new training methods. The paper presents a
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::0aee38baa385c3ffb67f77a7333e5968
https://doi.org/10.1109/hsi47298.2019.8942614
https://doi.org/10.1109/hsi47298.2019.8942614
Publikováno v:
2019 International Conference on Information and Digital Technologies (IDT).
Error Back Propagation algorithm is one of the most popular method for training artificial neural networks. Unfortunately, this is also one of the slowest due to constant and small learning rate parameter used to update weights of neurons. There are
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::3201c0e0cb6d6824c369d10a1e999e8b
https://doi.org/10.1109/dt.2019.8813440
https://doi.org/10.1109/dt.2019.8813440
Publikováno v:
HSI
Deep neural networks are able to solve much more complex and nonlinear problems than very popular but shallow technologies such as ELM, SVR or SLP. Despite of their power deep neural networks are difficult to apply due to problems with effective and
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::dba44165bd6f01a02c0c00d091777d85
https://doi.org/10.1109/hsi47298.2019.8942622
https://doi.org/10.1109/hsi47298.2019.8942622
Publikováno v:
2019 IEEE 23rd International Conference on Intelligent Engineering Systems (INES).
The paper presents a new method for improvement of the Error Back Propagation, one of the most popular algorithms for training artificial neural networks, that is based on the estimation of the learning rate by the approximation of the error of the o
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::b001ea5229c7a5a7c2be5e8fcb02fc28
https://doi.org/10.1109/ines46365.2019.9109445
https://doi.org/10.1109/ines46365.2019.9109445
Publikováno v:
Artificial Intelligence and Soft Computing ISBN: 9783030209117
ICAISC (1)
National Information Processing Institute
ICAISC (1)
National Information Processing Institute
The Fully Connected Cascade Networks (FCCN) were originally proposed along with the Cascade Correlation (CasCor) learning algorithm that having three main advantages over the Multilayer Perceptron (MLP): the structure of the network could be determin
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::b607509532227029d5ebc052e7e44dca
https://doi.org/10.1007/978-3-030-20912-4_23
https://doi.org/10.1007/978-3-030-20912-4_23
Publikováno v:
IECON
Deep learning become a popular trend in current research and applications. Deep neural networks are important part of this trend. The paper shows the effect of neural network architecture on its power and capacity for solving complex, nonlinear probl
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::eec4e7840f024690f0ab5f90e9a7bb73
https://doi.org/10.1109/iecon.2018.8591593
https://doi.org/10.1109/iecon.2018.8591593
Publikováno v:
Artificial Intelligence and Soft Computing ISBN: 9783319912523
ICAISC (1)
ICAISC (1)
The saturation of particular neuron and a whole neural network is one of the reasons for problems with training effectiveness. The paper shows neural network saturation analysis, proposes a method for detection of saturated neurons and its reduction
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::0fb489c328c20475bc56b3e2ed94a93d
https://doi.org/10.1007/978-3-319-91253-0_11
https://doi.org/10.1007/978-3-319-91253-0_11
Publikováno v:
Artificial Intelligence and Soft Computing ISBN: 9783319912523
ICAISC (1)
ICAISC (1)
Successful training of artificial neural networks depends primarily on used architecture and suitable algorithm that is able to train given network. During training process error for many patterns reach low level very fast while for other patterns re
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::f5be8b9cab68c077c51068007ccd62e9
https://doi.org/10.1007/978-3-319-91253-0_19
https://doi.org/10.1007/978-3-319-91253-0_19